Ionized physical vapor deposition is a technique used to deposit films, typically of metal or metal compounds, on semiconductor wafers during the fabrication of semiconductor devices. Ionized physical vapor deposition is advantageous for metallization of high aspect ratio vias and trenches extending into semiconductor wafers. In particular, ionized physical vapor deposition provides excellent sidewall and bottom coverage for vias and trenches.
In ionized physical vapor deposition, a metal target is mounted within a vacuum chamber and biased with a negative voltage relative to the grounded metal walls of the vacuum chamber. A sputtering gas, such as argon, is flowed into the vacuum chamber. The negative voltage on the target excites the sputtering gas proximate to the target into a plasma state and accelerates ions from the plasma to bombard an exposed surface of the target. The ion bombardment sputters metal atoms from the target that are ejected with a distribution of angular trajectories. A semiconductor wafer is held in the vacuum chamber near, and parallel to, the target. A relatively high-density plasma is generated in a region of the vacuum chamber between the target and the semiconductor wafer. The plasma ionizes a large proportion of the metal atoms ejected from the target, which are thermalized in the plasma. The semiconductor wafer is also negatively biased relative to the electrically grounded metal walls of the chamber, which accelerates the ionized sputtered target atoms from the plasma toward the semiconductor wafer. The target atoms impact the semiconductor wafer with more perpendicular trajectories than characteristic of non-ionized sputtering, which greatly increases the amount of material deposited at the bottom of high aspect ratio vias and trenches.
Etching of the metal film forming on the surface of the wafer, which occurs concurrently with the metal film's deposition, is an important component of the ionized physical vapor deposition process. Operating pressures of tens of mTorr are required in the vacuum chamber of the ionized physical vapor deposition apparatus to permit the material to thermalize in the plasma. At low chamber pressures (e.g., <30 mTorr), metal sputter etched from the surface of the semiconductor wafer has a large thermalization length (e.g., 100 mm to 200 mm). A portion of the sputter etched metal will return to the wafer surface and be reincorporated into the forming metal film as redeposited material. The rate of redeposition is influenced by the chamber pressure. At low chamber pressures, only a small fraction of the sputter etched metal will return back to the surface because of the large thermalization length. Hence, redeposition is negligible in comparison to the etch rate.
As the chamber pressure increases, the density of the various species in the plasma likewise increases. The mean-free-path of the metal atoms sputter etched from the metal film becomes small in comparison to the dimensions of the vacuum chamber, so that randomization occurs due to collisions and gas scattering of sputter etched metal. As a result, the thermalization length for the sputter etched metal drops. At chamber pressures greater than 30 mTorr, redeposition of the sputter etched metal has a significant impact on the etch rate of the metal film.
In particular, the thermalization length required for atoms sputtered from metal film at a kinetic energy of several electron volts to reach a kinetic energy on the order of 0.1 eV to 0.2 eV is on the order of 15 mm or less at these pressures. Measured etch rates are influenced by the redeposited sputter etched metal in a manner that cannot be reliably ascertained. Consequently, because of the significance of the redeposition, the actual etch rate at higher chamber pressures cannot be accurately evaluated in an ionized physical vapor deposition apparatus using conventional calibration wafers.
Moreover, the amount of redeposition experienced by the substrate surface as the metal film is deposited may exhibit a significant radial dependence relative to the azimuthal centerline of the substrate. In other words, a region of the substrate near the substrate edge may experience a reduced redeposition rate than a region of the substrate located near the substrate center. The radial dependence may change dramatically with a relatively minor change in the chamber pressure. As a result, the ability to measure the actual etch rate is not only impacted by the redeposition process but may also exhibit a dependence based upon the position on the substrate surface. Conventional calibration wafers are unable to evaluate this position dependent behavior of the actual etch rate in an ionized physical vapor deposition apparatus.
There is thus a need for a calibration wafer and method that can measure actual etch rates in an ionized physical vapor deposition apparatus at chamber pressures characterized by a short thermalization length.